Top Banner
Real-life Data Visualization Business Intelligence & Analytics INSY-442, Winter 2014 Michael Deutsch McGill Info Systems Resources (ISR)
58
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Page 1: Real-life Data Visualization - guest lecture for McGill INSY-442

Real-life Data Visualization

Business Intelligence & Analytics

INSY-442, Winter 2014

Michael Deutsch McGill Info Systems Resources (ISR)

Page 2: Real-life Data Visualization - guest lecture for McGill INSY-442

Agenda

Personal intro

Data visualization defined

What is good visualization?

Visualization in business

Exercise: Original analysis + visualization

Page 3: Real-life Data Visualization - guest lecture for McGill INSY-442

Profile: Mike Deutsch

Education: U. Penn (Philadelphia), 1998.

Computer Science & Cognitive Science

Management @ Wharton

15 years as a Designer/Analyst/Developer in

educational environments

McGill central IT since 2002:

Student systems (Banner/Minerva/uApply)

LMS (myCourses)

Data Warehouse & Business Intelligence

Page 4: Real-life Data Visualization - guest lecture for McGill INSY-442

Profile: Mike Deutsch

Some of my analyses over the years:

MyCourses

Traffic patterns

Pedagogical use

McGill Business Intelligence

Strategic Enrollment Management

Course Teaching patterns

Graduate capacity planning

Mercury course evaluations

Page 5: Real-life Data Visualization - guest lecture for McGill INSY-442

Necessary disclaimers

I strive for accuracy, but facts may not always

be exactly as reported.

All opinions expressed herein are my own,

and may not correspond to official University

positions on whatever it is I’m talking about.

Visualization is a blend of Art & Science;

Keep both hats on.

Please ask questions!

Page 6: Real-life Data Visualization - guest lecture for McGill INSY-442

Data visualization defined

Page 7: Real-life Data Visualization - guest lecture for McGill INSY-442

Data visualization defined

The pictorial representation of data.

How?

Via graphing tools or custom design.

Why?

To communicate information clearly and effectively

through graphical means.

In the right conditions, this sets in motion a narrative

in the viewer’s head.

Their brain can do what brains do best:

see patterns and relationships between dry bits of data.

Page 8: Real-life Data Visualization - guest lecture for McGill INSY-442

Data visualization defined

Visualization ≠ Numeric tables(this is only data)

Page 9: Real-life Data Visualization - guest lecture for McGill INSY-442

Data visualization defined

Visualization ≠ Illustrations or Diagrams(this is only pictorial)

Greek optics paper, 11th century CE

Darwin on speciation, 1837

Page 10: Real-life Data Visualization - guest lecture for McGill INSY-442

Data visualization defined

Visualization ≠

Graphic design

or “infographics” (this forces the story)

Megan Jaegerman,

New York Times, 1992

Page 11: Real-life Data Visualization - guest lecture for McGill INSY-442

Data visualization defined

To do it well: Data + pictorial

Functionality and aesthetics and go hand in hand.

The data itself tells the story.

Bonus: it’s backed up by emerging neuroscience.

Left brain/Right brain

Page 12: Real-life Data Visualization - guest lecture for McGill INSY-442

Data visualization defined

The pictorial representation of data:

Few, Stephen (2013):

Data Visualization for

Human Perception.

interaction-design.org

Page 13: Real-life Data Visualization - guest lecture for McGill INSY-442

Data visualization defined

Why is this important right now?

Explosion in visualization online & everywhere

Web analytics

Business Intelligence

Journalism (all forms)

Annual reports, presentations, etc.

As Management students & professionals,

there are things YOU can do to improve the charts &

tables you produce. Your work can become more

informative and influential.

Page 14: Real-life Data Visualization - guest lecture for McGill INSY-442

What is good visualization?

Page 15: Real-life Data Visualization - guest lecture for McGill INSY-442

What is good visualization?

Recap: You can’t just “visually represent data”. You have to somehow make it good.

~~~~

In the right conditions, this sets in motion a narrativein the viewer’s head.

Their brain can do what brains do best: see patterns and relationships between dry bits of data.

~~~~

To do it well:

Data + pictorial

Functionality and aesthetics and go hand in hand.

The data itself tells the story.

Page 16: Real-life Data Visualization - guest lecture for McGill INSY-442

What is good visualization?

The most exciting phrase to hear in science,

the one that heralds new discoveries,

is not “Eureka” (I found it) but “That's funny...”

Isaac Asimov (1920–1992)

Page 17: Real-life Data Visualization - guest lecture for McGill INSY-442

What is good visualization?

So, some examples…

Page 18: Real-life Data Visualization - guest lecture for McGill INSY-442

What is good visualization?

The original:

William

Playfair,

Scotland, 1786 “…the first quantitative

graphical form that did

not locate data either in

space, as had

coordinates and tables,

or time, as had

Priestley's timelines. It

constitutes a pure

solution to the problem of

discrete quantitative

comparison”

Wikipedia: William Playfair

Page 19: Real-life Data Visualization - guest lecture for McGill INSY-442

What is good visualization?

A great first example: GraphJam

Taking notes in class

When I become a professor

http://cheezburger.com/FlashBuilder/GraphJam

A little more serious: Information is Beautiful

http://www.informationisbeautiful.net/visualizations/mount

ains-out-of-molehills/

Page 20: Real-life Data Visualization - guest lecture for McGill INSY-442

What is good visualization?

Good news: You DON’T have to be a graphic artist.

Tools help.

Google Analytics, Google Charts, Google Maps

Excel

Tableau

And there are rules.

Page 21: Real-life Data Visualization - guest lecture for McGill INSY-442

What is good visualization?

THE RULES:

Show the data

Show restraint with everything that’s NOT data

Encourage comparison

Avoid distorting the data

Page 22: Real-life Data Visualization - guest lecture for McGill INSY-442

What is good visualization?

Show the data

(Duh)

But clearly

And show enough of it

Page 23: Real-life Data Visualization - guest lecture for McGill INSY-442

What is good visualization?

Show restraint with everything that’s NOT data

Simplify and subdue colors

“Calm” or remove structure: lines, grids, tick marks, labels

Page 24: Real-life Data Visualization - guest lecture for McGill INSY-442

What is good visualization?

Encourage comparison

Side-by-side or contrasting appearance

Also:

The trustworthiness

of beards

Page 25: Real-life Data Visualization - guest lecture for McGill INSY-442

What is good visualization?

Avoid distorting the data

3D effects and free-form visualization are risky

Page 26: Real-life Data Visualization - guest lecture for McGill INSY-442

What is good visualization?

Excel, pre-2007

Page 27: Real-life Data Visualization - guest lecture for McGill INSY-442

What is good visualization?

So who does it well in the business world? National Post:

http://news.nationalpost.com/category/graphics/

The Economist:

http://www.economist.com/blogs/graphicdetail

New York Times:

http://www.nytimes.com/interactive/2012/12/30/multimedia/2012-the-year-in-graphics.html?_r=0

Wall Street Journal:

http://online.wsj.com/public/page/news-interactive-features-trends.html

Tableau Public

https://www.tableausoftware.com/public/community/in-the-wild

Page 28: Real-life Data Visualization - guest lecture for McGill INSY-442

What is good visualization?

More at National Post:

2011 Canadian census

Immigration patterns

So many ways to die

Ontario budget breakdown (A misfire?)

Charts Bin (geog)

http://www.Chartsbin.com

Tableau – Public data sets, make your own.

(download)

http://www.tableausoftware.com/products/public

Page 29: Real-life Data Visualization - guest lecture for McGill INSY-442

What is good visualization?

Geog charts

http://cryscresc.net/current-worldwide-annual-coffee-

consumption-p-0 -- original source is http://chartsbin.com/

NYT croissants vs. white bread

Gazette population flow

Add time dimension

http://chartsbin.com/view/1016 - current worldwide alcohol

consumption

http://chartsbin.com/view/1026 - Change over last 5 years

What can you NOW see? Suit your charts to your

audience and what you’re trying to convey.

Page 30: Real-life Data Visualization - guest lecture for McGill INSY-442

What is good visualization?

Jacques Bertin Edward Tufte

1918 -2010

French cartographer

Semiologie Graphique (1965)

pioneered a graphic

vocabulary

American social scientist &

statistician

The Visual Display of

Quantitative Information

(1967)

edwardtufte.com

Page 31: Real-life Data Visualization - guest lecture for McGill INSY-442

What is good visualization?

Stephen Few

British programmer &

designer

Social data sets &

interactive visualization

InformationIs Beautiful.net

American BI/DWH

innovator

Information Dashboard

Design (2006) led the

modern business

dashboard era

PerceptualEdge.com

David McCandless

Page 32: Real-life Data Visualization - guest lecture for McGill INSY-442

Visualization in business

Page 33: Real-life Data Visualization - guest lecture for McGill INSY-442

Visualization in Business

We saw the “Rules” of data viz:

Show the data

Restraint in everything that’s NOT data

Encourage comparison

Avoid distorting the data

Let’s see them in use in the business world…

Can occur in different contexts, on different

platforms.

The goal in business is always:

viz narrative action.

Page 34: Real-life Data Visualization - guest lecture for McGill INSY-442

Visualization in BusinessData

viz

B.I.

Analytic

s

Data

mining

Page 35: Real-life Data Visualization - guest lecture for McGill INSY-442

Visualization in Business

Sun Life Financial Inc.:

Large multinational, financial services, Toronto HQ.

[Annual Report of Sun Life Financial: 2010 & 2012]

Page 36: Real-life Data Visualization - guest lecture for McGill INSY-442

Visualization in Business

McGill University:

Large public educational institution.

Mission: Teaching, Scholarship, Service.

Just like corporations, McGill analyzes and solves:

“Production”

Operations

Budgeting

Marketing

Competition

Regulation

Page 37: Real-life Data Visualization - guest lecture for McGill INSY-442

Visualization in Business

[Corporate structure]

[Early reports & dashboards]

Page 38: Real-life Data Visualization - guest lecture for McGill INSY-442

Visualization in Business

Before: http://www.mcgill.ca/es/registration-statistics

Page 39: Real-life Data Visualization - guest lecture for McGill INSY-442

Visualization in Business

1st gen Enrolment Analysis dashboard: ProClarity

Page 40: Real-life Data Visualization - guest lecture for McGill INSY-442

Visualization in Business

2nd gen dashboard: Microsoft PerformancePoint

Minimalist look, parallel design rules 1 & 2

+ Immigration, Fee residence, Admit region

Filters for the dimensions ppl use in this context

Interactive tool allows drill down/across/through

In-browser, pop out to Excel

Page 41: Real-life Data Visualization - guest lecture for McGill INSY-442

Visualization in Business

Page 42: Real-life Data Visualization - guest lecture for McGill INSY-442

Visualization in Business

[Principal’s Report – Fact Book]

[PIA – KPI Report to Senate]

[Provost’s Budget Book]

Page 43: Real-life Data Visualization - guest lecture for McGill INSY-442

Conclusion

Page 44: Real-life Data Visualization - guest lecture for McGill INSY-442

Conclusion

Good viz allows people’s eyes & brain to mine

data.

Presenting the data is just the first step.

To really have impact, you have to present it well.

Data viz is a mix of Art & Science, but there are

tools and techniques you can use.

Yes, it makes a BIG difference in business.

Page 45: Real-life Data Visualization - guest lecture for McGill INSY-442

THANK YOU.

Contact:

[email protected]

Page 46: Real-life Data Visualization - guest lecture for McGill INSY-442

Exercise:

University Research Analysis

Page 47: Real-life Data Visualization - guest lecture for McGill INSY-442

Exercise: University Research Analysis

Introduce the business area

Produce original analysis from SSHRC awards data

Push your visualization

Show us what you came up with

Page 48: Real-life Data Visualization - guest lecture for McGill INSY-442

SSHRC: In Few Words

The Social Sciences and Humanities Research Council (SSHRC) is

the federal agency that promotes and supports postsecondary-based

research and training in the humanities and social sciences.

Facts and Figures

Full-time professors: 22,410

Full-time doctoral students: 19,005

Full-time master’s students: 43,150

SSHRC funding (2010-11)

Number of applications received: 12,925

Number of experts involved in peer review: 6,748

New grants, fellowships and scholarships awarded: 4,381

New partnerships with industry, government and NGOs: 118

Value of all grants, fellowships and scholarships supported: $332.4 million

Page 49: Real-life Data Visualization - guest lecture for McGill INSY-442

# of

projects

Amount

($ millions)

% of

funding

Environmental science and technologies 475 $21.7 7%

Natural resources and energy 43 $2.6 1%

Health and related life sciences and

technologies370 $19.0 6%

Information and communications technology 658 $23.6 7%

Management, business and finance 695 $20.6 6%

Arts and culture 1,875 $52.2 16%

Economy, employment and markets 717 $26.7 8%

Education and learning 793 $27.3 8%

Human development 661 $21.9 7%

Social development 1,855 $64.4 19%

Governance of society and institutions 1,092 $35.4 11%

Other areas of research 539 $16.9 5%

TOTAL: 9,773 $332.4 100%

Page 50: Real-life Data Visualization - guest lecture for McGill INSY-442

Exercise: University Research Analysis

Pick up the XLS 2010 file.

Produce original analysis from SSHRC awards data Approach ONE of the questions on the XLS tab.

Produce a pivot table or chart.

Create a DAX expression: distinct # researchers, avg $, etc.

Use THE RULES: Show the data

Restraint in everything that’s NOT data

Encourage comparison

Avoid distorting the data

Display in different (appropriate) ways

Attend to what’s important

Page 51: Real-life Data Visualization - guest lecture for McGill INSY-442

Exercise: University Research Analysis

Push your viz:

Color schemes

Chart types (independent vs. stack vs. 100% stack)

Conditional formatting (heatmap, bars, etc.), sparklines

Page 52: Real-life Data Visualization - guest lecture for McGill INSY-442

Exercise: University Research Analysis

What did you come up with?

Page 53: Real-life Data Visualization - guest lecture for McGill INSY-442

Appendix:

Visualization Resources

Page 54: Real-life Data Visualization - guest lecture for McGill INSY-442

Appendix:

Other fun visualizations

Page 55: Real-life Data Visualization - guest lecture for McGill INSY-442

Old school:

The newspaper/magazine “infographic” style:http://www.theonion.com/articles/worst-parts-of-our-jobs,7433/

Page 56: Real-life Data Visualization - guest lecture for McGill INSY-442

Infographic renaissance…

http://graphjam.memebase.com

Page 57: Real-life Data Visualization - guest lecture for McGill INSY-442

Infographic renaissance…

http://poorlydrawnlines.com/

http://ilovecharts.tumblr.com/

Data visualization

http://popchartlab.com/

http://www.informationisbeautiful.net/

Page 58: Real-life Data Visualization - guest lecture for McGill INSY-442

XKCD

Self-description

Money

Height

Numbers